比较与tf.nn.moments的功能差异

tf.nn.moments

tf.nn.moments(x, axes, shift=None, keepdims=False, name=None) -> Tensor

更多内容详见tf.nn.moments。

mindspore.nn.Moments

class mindspore.nn.Moments(axis=None, keep_dims=None)(x) -> Tensor

更多内容详见mindspore.nn.Moments。

差异对比

TensorFlow：沿指定轴axes计算输入x的均值和方差。

MindSpore：MindSpore此API实现功能与TensorFlow一致。

分类	子类	TensorFlow	MindSpore	差异
参数	参数1	x	x	-
	参数2	axes	axis	功能一致，参数名不同，TensorFlow中该参数无默认值，MindSpore中该参数默认值为None
	参数3	shift	-	TensorFlow的该参数在当前实现中未使用，是无用参数。MindSpore无此参数
	参数4	keepdims	keep_dims	功能一致，参数名不同
	参数5	name	-	不涉及

代码示例1

两API用于计算Tensor指定轴的均值和方差，用法相同。

# TensorFlow
import tensorflow as tf
import numpy as np

x = tf.constant(np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]]), dtype='float32')
mean, variance = tf.nn.moments(x, axes=0, keepdims=True)
print(mean.numpy())
# [[[3. 4.]
#   [5. 6.]]]
print(variance.numpy())
# [[[4. 4.]
#   [4. 4.]]]

# MindSpore
import mindspore
from mindspore import Tensor
from mindspore import nn
import numpy as np

x = Tensor(np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]]), mindspore.float32)
net = nn.Moments(axis=0, keep_dims=True)
mean, variance = net(x)
print(mean)
# [[[3. 4.]
#   [5. 6.]]]
print(variance)
# [[[4. 4.]
#   [4. 4.]]]

代码示例2

两API用于计算Tensor所有值的均值和方差，TensorFlow的API需要手动指定所有轴，MindSpore的API默认指定所有轴。

# TensorFlow
import tensorflow as tf
import numpy as np

x = tf.constant(np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]]), dtype='float32')
mean, variance = tf.nn.moments(x, axes=[0, 1, 2])
print(mean.numpy())
# 4.5
print(variance.numpy())
# 5.25

# MindSpore
import mindspore
from mindspore import Tensor
from mindspore import nn
import numpy as np

x = Tensor(np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]]), mindspore.float32)
net = nn.Moments()
mean, variance = net(x)
print(mean)
# 4.5
print(variance)
# 5.25